Dave Haynie
Forum Replies Created
-
I second that.. I’ve been playing around with the improved Main Concept AVC, and it’s very good. Plus, it does VBR, which is sometimes what you want.
I have typically used the ~8Mb/s AVC for uploading of 720p material to YouTube and other online sites. They’ll naturally transcode it to something lesser, but I haven’t found I get better results on most material going higher.
It’s also kind of interesting to see the kind of bitrates consumers are not typically viewing on the various services. Netflix HD, for example, is VC-1 720p24 or 720p30 at up to 3.6Mb/s. Apple’s newly launched 1080p24 option (which is chewing though the $50/month LTE accounts on the iPad “3” in 3 hours) is somewhere in the 3.8-4.0 Mb/s range, AVC of course… and I’ve read a few accounts (by Apple fans, no doubt) comparing this encoding of films to the same BD favorably enough (eg, they’ll give the nod to the BD, but clearly, the HD edition is far superior to DVD or 720p versions).
So, in short, for most consumer viewing purposes, doubling that for half the resolution (eg, 720p at 8Mb/s) is probably just dandy for online stuff. 720p24 or 720p30 may also be a “safer” format to just hand to a consumer, since this ought to play well on most modern PCs. 1080p is going to stress a 2-core laptop if you don’t have GPU acceleration (which Windows 7’s built-in AVC CODEC will use… though most laptops don’t have a capable enough GPU to offer such help). And forget it on most tablets… if you know it’s destined for one, choosing an iPod/iPhone/iPad is a pretty safe bet (iOS is a popular platform, sure, but most Android devices have similar or better video playback chops, at least those not running Cyangen(mod) or other “community” builds of Android).
-Dave
-
[Gilles Gagnon] “2. I’m told USB3 is faster than Esata. Would I not be better off to go USB3 (with sata drives in dock)?”
The peak data rate on USB 3.0 is 5Gb/s in “SuperSpeed” mode. The preak data rate on eSATA will be 1.5Gb/s, 3.0Gb/s, or 6.0Gb/s, depending on the version of SATA you have (people say “SATA-I”, “SATA-II”, etc. but that’s not the official designation of the different editions).
The fastest HDD I know of are the enterprise-level SSD drives, which don’t write terribly fast, but hit 150-160MB/s sustained reads, if things go well. 160MB/s = 1.3Gb/s. So basically, you shouldn’t care all that much about the interface, when it comes to speed; all the latest interfaces (including Thunderbolt, if you want to toss that into the mix, and even GigE is a contender) are fast enough for a single drive. In fact, it’s common to drive a whole RAID or JBOD box from a single eSATA port, and you probably still get full speed (which is kind of the point of 6Gb/s SATA… there’s no possible reason you need that speed for a single drive, yet).
USB 3.0 has the advantage of still working with USB 2.0 gear, which is useful if you have a reason to move the drive from PC to PC. Though I have the same SATA Frames in both my desktop PCs… and GigE networking between them as well, so moving files around isn’t much of an issue.
-Dave
-
None of the above… exactly. I use internal SATA drives, one per major project, AS replaceable drives, using a SATA frame like this one: https://www.amazon.com/gp/product/B0028Y4DAO/ref=oh_o00_s02_i01_details
So you get full SATA speed, no need for an external power supply (though I do have external RAID and a NAS, they’re usually used for other things). And when I’m done, the drive goes on the shelf as an archive of the finished project, in case I need to go back to it at some point. Sure beats backing up to a stack of BDs…
-Dave
-
[Brian Mason] “When I render the video to the hard drive I select the Video for Windows (*.avi) which results in NTSC DV with
Video: 29.970 fps, 720X480Is there a better format to use?”
So you’re capturing over Firewire, yielding a DV AVI. You edit, and render back to a DV AVI. The only thing that’s likely to change input to output are changes you make in Vegas. Vegas will “Smart Render” DV to DV — it makes an exact digital copy, input to output, for any video not being edited. So if you’re color grading or making other adjustments, it all gets re-rendered. If you’re just cut editing, it’s virtually identical to the input.
Do make sure you’re rendering to DV. There are dozens of types of video that can be put in an AVI wrapper, DV is just one of them. They vary widely in quality.
You can sanity check youself, too. Grab an unused tape, print your edited video to back to the Sony, and play it on your TV as you have in the past. If it looks bad, that’s an indication that you’re damaging it somehow on that trip through Vegas. Large adjustments of brightness and contrast, some other effects, can do that — video can be fairly fragile, like anything that’s only 24-bit color.
If it looks as good as the original on your TV, then it’s some aspect of your computer playback that’s the issue.
[Brian Mason] “I have a AOC back lit LED monitor that claims 50,000,000:1 DCR. It is a 1900 X 1200 and I use a DVI connection.”
Look like AOC makes monitors that sound fairly good… I didn’t find one that’s both LED backlit (an “LED” monitor is a marketing term for an LCD with a LED backlight… “back lit” is superior to “edge lit”… it’s modulation of the LEDs in small zones that gives you the crazy dynamic contrast range) and 1920×1200, but I did find a 1920×1080 IPS panel unit.
Anyway, with the right software behind it, this should look similar to an HDTV of the same size, when playing video… most monitors these days are borrowing heavily from TV technology — thus the 1920×1080 rather than 1920×1200. But a television does a bunch of high quality upscaling and de-interlacing for video playback, and the video is always in full sync with the screen. That’s not necessarily the case with a computer display.
Anyway, give the print-to-tape exercise a go, and that will at least split the problem for you.
-Dave
-
[Gregory Bauhof] “I need a some suggestions as to what is the best format to render into in order to make a DVD.”
If you want a real DVD-Video format DVD, you have one choice: MPEG-2. Aside from that, you want the rate to match your source video.
[Gregory Bauhof] “I shot a film with my Sony HVR-V1U on the 24p setting. After editing in Vegas Pro 11, I’ve tried to render it into a format that would be good for DVDs in order to send out to film festivals. But if I render using the 24p setting, I get blurs whenever people move quickly. It’s a short sci-fi film so there are moments of a lot of motion.
“Check what Vegas thinks of your V1U video… there’s a pretty good chance it think you have 60i video there, not 24p. One of the issues with tape formats is that tapes always run the same speed, and thus, the video CODECs (like DV or HDV) always run the same speed. Because of this, back in the DV days, when someone got the great idea of adding 24p, they didn’t change the tape format, they simply recorded 24p as 60i, with 3:2 pulldown added (telecine). In order to correctly edit this, you have to detect the telecine and invert it (eg, simply telling Vegas to render to 24p isn’t sufficient).
But check you clips in Vegas… see if Vegas thinks they’re 24p or 60i. Is the output video showing actual blur, or something worse? See, the problem gets down to a number of things related to Vegas’ handing of interlaced vs. progressive video during scaling. When you convert 1080/24p to 480/24p for DVD, it should be pretty smooth. However, if Vegas thinks you’re 60i, it’s going to want to de-interlace (which, of course, is in the 2:3:3:2 or whatever telecined cadance of frames to fields), then downconvert… much worse looking than just the downconversion you’re get 24p -> 24p. For downsizing, you probably want the project set to “best”, which will use BiCubic rather than BiLinear interpolation.
As far as blurring goes… 24p and MPEG-2 are both limited in their ability to deal with motion. Whether you get motion blur in your video is in part based on your shutter speed, of course. But heavy motion will also result in more MPEG-2 artifacts in the original video.
-Dave
-
The camera you’re using, Sony DCR-TRV130, is a Digital 8 camcorder. Assuming you’re bring the video in via Firewire, while the process seems like video capture, you’re actually doing a digital copy from the camcorder. What you capture on the PC will be exactly what’s on that tape, same as what you see on the TV.
Are you copying the DV AVIs directly to the server, or converting them? That’s one source of quality loss… converting the name format to something else can certainly damage the quality.
It may just be the difference in viewing. When you watch these on a television, you’re viewing a video shot in PAL or NTSC native format, on a device presumably designed to display that format optimally, despite its relatively low resolution. When viewed on a PC, you’re expanding that 720×480 or 720×576 to a full PC screen… yours may not be 1920×1200 like mine, but chances are, the PC screen is much higher native resolution. And you’re viewing much closer. The two of these conspire to make SD video look much worse than on TV.
Next come interlacing… that camera only shoots in 60i (NTSC) or 50i (PAL), which looks ok on an old timey CRT, and can look even better if viewed on an HDTV with a good upscaler/deinterlacer. But it can look pretty grim on a computer monitor.
It could also be your monitor. A CRT delivers about a 15,000:1 contrast ratio, more or less. A modern LCD/LED television can kind of simulate this, and even more; other technologies, like DLP or Plasma, exceed the contrast ratio of a CRT. But PC LCDs… not so much. You may have something like 800:1 or 1000:1… not much dynamic range. Worse yet, if you have a TN LCD rather than an MVA/PVA or IPS display, you may only have 6-bits per color resolution, rather than 8-bits per color.. so you’ll see banding and other low-color effects that you don’t see on TV.
It could also be your player. Image processing of an SD video will enhance viewing on a computer monitor just as it would on an HDTV. Simple programs like VLC or Windows Media Player don’t do much here; you may want to try “Splash Lite”, which does some image processing you won’t find in these other two.
-Dave
-
[Matt Whittemore] “I’ve run memory tests in OSX and it comes out fine.”
Most memory tests are not written well enough to isolate anything but a gross failure in a memory device. I was designing memory systems for the better part of 11 years at Commodore, and the memory test we put together in the lab (unfortunately lost) did things I’ve not seen in other tests. Like emulating the behavior of code, for example.
[Matt Whittemore] “OSX has never crashed and works flawlessly except when running FCPX”
If you’re stressing the system under OSX in other ways (eg, not just with FCPX), that’s the best sign that it’s not your hardware. FCPX is new enough to certainly be suspect.
The primary software issue that’ll see a modern OS crashing regular are bad drivers. You definitely want to ensure you have the latest… and don’t count on Windows Update to find them. It’s certainly going to update the Microsoft written components, but Microsoft’s policies and history has pretty much worked against their being supplied with recent drivers. Now, of course, if you have Mac-exclusive hardware of any kind, you may be SOL trying to find recent drivers. Particularly if it’s up to Apple to provide them for Windows; in the Windows ecosystem, system OEMs like Apple are usually responsible for providing the drivers, and it wouldn’t exactly be against Apple’s interest to keep alive the meme that Windows is crashy (it has been at various times, but Windows 7 is quite stable these days… the PC I’m typing this on, my main machine, seems to have been up for 18 days, 6 hours, 13 minutes, and 54 seconds as I write this). Not to suggest a conspiracy or anything.
But I would expect most Macs to be pretty bog standard Intel-based PCs these days. Intel’s actually been pretty good about driver updates, and you can go to Intel.com and get a scanning program that’ll find the drivers they recommend as updates. Graphics drivers are often buggy in weird ways; I’m running recent ATi drivers myself, and the system is quite stable.
If you do get a BSOD, try to write down some of the on-screen info, then do a search on it — you may be seeing a well-known error.
-Dave
-
[Jim Greene] “OK, to piggy-back off this topic, how do I render correctly for DVDA a 24p video with aspect ratio of 1.85:1. I can render easily using Sony AVC by simply changing the framesize to 1980×816, or 1280×544.”
You render to 1920×1080 or 1280×720, assuming you want to make a Blu-ray. Like DVD, BD only supports the two formats: 4:3 or 16:9. You can add letterbox bars in Vegas (well, you’ll get them if you have 1.85:1 source material), or you can render out without the bars, and let DVDA add them… which means it has to re-render.
-Dave
-
[Phil Peacock] “Hmmm. My conclusion so far too. Odd though, when you think about it, not having the convenience of skipping around in a video compilation (like ‘scene selection’ or even ‘extras’) especially when hyperlinks are so prevalent elsewhere.”
You have this, but only in formal delivery formats. MPEG-4 isn’t a formal deliver format, it’s Yet Another Motion Picture Experts Group media format, which really just concentrates on building useful pieces that other people will use to build on top of. So we have AVCHD, Blu-ray, DivX, and maybe a few other formats that offer some degree of DVD-like navigation, all based on MPEG-4 standards, but building beyond that. MPEG-4 also drives satellite TV here in the states, and second-generation digital TV in Europe… neither of which have any need for navigation…. they do non-interactive broadcast streaming.
I might have expected something more along these lines from Microsoft, given they had all their own standards. But they mimiced the MPEG folks… they actually did release an interactive film format based on WMV, dubbed WMV/HD (very creative name), which was high definition, usually 720p, on red laser DVDs. They used HTML and Javascript to author the disc — they assumed a player that knew how to respond to track offsets being sent.
-Dave
-
[John Rofrano] “Today we rarely use the term Anamorphic Widescreen anymore. We just assume that Widescreen is full 16:9. “
Yup. Historically, widescreen was anamorphic. You fit an anamorphic lens to you film camera.. this is a non-spherical lens, which is essentially a normal lens vertically, a wide-angle horizontally, so it optically “compresses” the image. You have to fit an anamorphic projection lens to the projector on play back. Prior to this, they did wide-screen by cropping, so much of the potential image was lost.
DVDs do the same things… in a way. There’s no actual “squeeze” involved, they just define the pixels as rectangular. In fact, they’re also rectangular for SD, just more exaggerated for HD.
[John Rofrano] “Which is what I said. You get more horizontal resolution. As I pointed out earlier, you were probably confused by them talking about non-anamorphic widescreen having only 345 lines of vertical resolution becasue of the letterboxing. “
That’s just like the cropping done way back in the film world. You can crop a 16:9 image out of a 4:3 SD image, and most HDTVs will actually blow that up to 16:9. But you’re losing that resolution. Better to just define a 16:9 format — so they did, but using the same pixel count.
These same issues are likely to crop up as we move beyond HD…
-Dave